EP3614220A1 - Détermination des états d'un dispositif au moyen des machines à vecteur de support - Google Patents

Détermination des états d'un dispositif au moyen des machines à vecteur de support Download PDF

Info

Publication number
EP3614220A1
EP3614220A1 EP18189722.4A EP18189722A EP3614220A1 EP 3614220 A1 EP3614220 A1 EP 3614220A1 EP 18189722 A EP18189722 A EP 18189722A EP 3614220 A1 EP3614220 A1 EP 3614220A1
Authority
EP
European Patent Office
Prior art keywords
classification
state
indicates
operating
classification volume
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP18189722.4A
Other languages
German (de)
English (en)
Inventor
Jonas Deichmann
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Siemens AG
Original Assignee
Siemens AG
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Siemens AG filed Critical Siemens AG
Priority to EP18189722.4A priority Critical patent/EP3614220A1/fr
Priority to EP19761763.2A priority patent/EP3811170B1/fr
Priority to CN201980054387.XA priority patent/CN112585548B/zh
Priority to PCT/EP2019/071895 priority patent/WO2020038815A1/fr
Priority to ES19761763T priority patent/ES2927838T3/es
Priority to US17/270,000 priority patent/US11244250B2/en
Publication of EP3614220A1 publication Critical patent/EP3614220A1/fr
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
    • G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
    • G05B23/0275—Fault isolation and identification, e.g. classify fault; estimate cause or root of failure
    • G05B23/0281—Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models

Definitions

  • Modern devices are often equipped with a large number of sensors which can detect a wide variety of operating parameters of the device before, after or during its operation.
  • Certain constellations of operating parameters may indicate that a device is in a normal state or, conversely, indicate that the device is in a fault state.
  • support vector machines From the area of machine learning (English “machine learning”) so-called support vector machines, SVM (English “support vector machine”) are known, which are sometimes also referred to as “support vector machine”.
  • SVM Support vector machine
  • Such support vector machines are trained with training data in a parameter space in order to generate boundaries ("class boundaries” or “classification boundaries”) between different classification volumes of the parameter space in such a way that the widest possible area around these boundaries remains as free as possible from the training data.
  • the generated class boundary can be used to determine in which classification volume a certain point, in particular the support vector machine, previously unknown, is positioned in the parameter space. In this way, support vector machines enable classification of n-dimensional points.
  • the computing device can in particular be a processor with an associated memory, a cloud computing platform, a microcontroller, an application-specific integrated circuit, ASIC, and / or the like.
  • the detection device can in particular comprise a plurality of sensors, each sensor for detecting at least one Operating parameters of the device is formed during its operation.
  • an electrical current can be detected by an electrical current sensor, a temperature by a temperature sensor, an electrical voltage by an electrical voltage sensor, etc.
  • the output module can be designed to generate a temporal trajectory of the operating point generated by the operating point module during the operation of the device and to determine or estimate an expected continuation of this trajectory in the future. Based on this expected trajectory, the output module can determine, in some advantageous embodiments, at which point in time the operating point of the device has arrived along the trajectory in a position that is or is to be assigned to a classification volume that indicates an error state of the device.
  • the output module can also be designed to indicate a point in time and / or a remaining time at which or after which the device is likely to be in the fault state. In this way, a user is supported by a continuous human-machine interaction not only to monitor the actual state of the device, but also to predict and possibly avoid a possible future fault state.
  • the support vector machine is preferably trained using labeled operating points in the n-dimensional operating parameter space, the labels indicating which classification volume the corresponding operating parameter point is to be assigned to and which group (for example normal state, fault state 1, fault state 2, etc.) is.
  • the invention provides a device, in particular an electrical machine, with a system according to the first aspect of the present invention.
  • the electrical machine can be, for example, an electric motor and / or an electrical generator act.
  • the device can also be, for example, a pump or a drive train or comprise a pump or a drive train.
  • the invention provides a computer program product which contains executable program code which is designed such that the program code, when it is executed, realizes the operating point module, the trained support vector machine and the output module.
  • the invention provides a non-volatile computer-readable storage medium which contains executable program code which is designed such that the program code, when executed, realizes the operating point module, the trained support vector machine and the output module.
  • the device can be an electrical machine, which is determined in particular by the operating parameters electrical voltage, electrical current, rotational acceleration and / or rotational speed can be characterized.
  • the trained support vector machine is designed to subdivide the n-dimensional operating parameter space into at least three classification volumes, the first classification volume indicating the normal state of the device and the second classification volume and a third classification volume different error states in the device index. In this way, an existing or predicted error state can be determined and identified more precisely.
  • the trained support vector machine separates the classification volumes from one another by two-dimensional levels.
  • the classification volumes are separated from one another by the trained support vector machine by so-called hyperplanes.
  • the support vector machine is configured to use a linear kernel.
  • vector-valued functions are often used ⁇ x ⁇ used to input vectors x ⁇ . which represent the operating parameters in the operating parameter space in a usually higher-dimensional feature space in which the separation by the support vector machine is usually easier.
  • a solution to the problem of finding class boundaries can then be found by setting up the Karush-Kuhn-Tucker conditions.
  • This can be advantageous instead of using the mapping function ⁇ x ⁇ to work explicitly, now advantageously with the kernel for determining the levels or hyperplanes separating the classification volumes instead of with the mapping function K x ⁇ . x ⁇ ' be worked.
  • Karush-Kuhn-Tucker conditions can then be set up using the kernel.
  • This technique after which K x ⁇ . x ⁇ ' when training the support vector machine and also during the classification instead of the mapping function ⁇ x ⁇ is used is called kernel trick.
  • the feature space has the same dimensionality as the original operating parameter space.
  • the detection device is set up to detect the detected operating parameters as portions (for example fractions, fractions or percentage values) of a respective corresponding maximum operating parameter value.
  • portions for example fractions, fractions or percentage values
  • normalization of the operating parameters can advantageously be carried out.
  • this allows the value of the individual operating parameters to be determined based on the classification limits of the classification volumes (that is to say on the levels or hyperplanes), from which conclusions can be drawn about the causes of errors.
  • This largest entry in terms of amount of the respective normal vector thus indicates a change in which parameter most likely leads to the operating point being from a position within the first classification volume (normal state) into the respective classification volume associated with the normal vector, which indicates an error state, emotional.
  • the method according to the second aspect of the present invention comprises the following features and steps:
  • the n-dimensional operating parameter space is divided into at least three classification volumes by the trained support vector machine, the first classification volume indicating the normal state of the device and wherein the second classification volume and a third classification volume indicate different error states of the device.
  • the operating parameters recorded can in particular, as explained in the foregoing, be recorded as portions of a corresponding maximum operating parameter value.
  • a respective normal vector can advantageously be determined on each level or hyperplane, which is the first Separates classification volume, which identifies the normal state of the device, from a classification volume, which indicates an error state of the device.
  • the value of the largest entry in terms of amount of each particular normal vector is also advantageously determined and output. As explained in the foregoing, it can be determined in this way which operating parameter is particularly important for reaching the respective fault condition.
  • the device can be an electrical machine, more precisely an electrical motor, which is designed, for example, with at least one vibration sensor and at least one temperature sensor.
  • Two or more vibration sensors can be arranged at different locations on the electric motor and / or two or more temperature sensors can be located at different locations on the electric motor.
  • Each measured value of each sensor of the electric motor can fill an entry (index) in an input vector.
  • the current input vector thus represents the current operating point of the electric motor.
  • the described technology can be used to determine at which point in the motor the problem occurred and whether the possible cause is more of a problem Overheating (greater weighting of the temperature sensor data) or excessive vibration (greater weighting of the vibration sensor data) was.
  • the device can be a pump, for example for waste water, which normally requires a certain amount of electrical current (first operating parameter) and a corresponding amount of electrical voltage (second operating parameter).
  • first operating parameter normally requires a certain amount of electrical current
  • second operating parameter a corresponding amount of electrical voltage
  • third operating parameter can also be measured.
  • the device can be an electric saw, for example a circular saw for woodworking.
  • An engine temperature, vibration and rotational speed can be measured and recorded as operating parameters.
  • a high engine temperature may indicate general overload or wear of the engine as the cause; strong vibration in combination with high temperature can indicate wear of the saw blade as the cause; and a slow rotation speed in combination with low vibration can indicate the reason that a wood to be sawn is too hard.
  • the device may include a vehicle driveline, e.g. an electric vehicle or a hybrid vehicle.
  • a current speed can be measured as the first operating parameter and a current torque as the second operating parameter. For example, if the drive train is stopped due to the exceeding of a maximum permissible torque, e.g. determine whether the cause is an excessive load on the axis or rather an excessive acceleration.
  • an ordered list of the entries of each specific normal vector can be determined and output according to their size (in each case in connection with their index value, so that it can be traced which entry belongs to which operating parameter) in order to provide a hierarchy of the importance of the corresponding operating parameters for reaching them of the respective fault condition. This can be of great advantage when evaluating the possible causes of errors, for example in a workshop.
  • Fig. 1 shows a schematic block diagram of a system 100 for determining a state of a device 200.
  • the device 200 can in particular be an electrical machine, for example an electrical motor and / or generator.
  • the device 200 can also be a pump or a drive train of a vehicle, for example an electric vehicle or a hybrid vehicle.
  • device 200 may include system 100 for determining the state of device 200.
  • a vehicle can also be provided which comprises both the system 100 and the system 200.
  • the system 100 comprises a detection device 10, which is designed to detect at least two operating parameters of the device 200 during its operation.
  • the detection device can, for example, each comprise a sensor 11, 12, which is each configured to detect the respective operating parameter.
  • the operating parameters are recorded as portions of a respective maximum operating parameter value.
  • the single ones Sensors 11, 12 can, for example, be set up in such a way that they already output the measured values as such portions.
  • the detection device 10 can also comprise a conversion unit, for example implemented by a processor, microcontroller, an ASIC or an FPGA, which is designed to convert the raw sensor data detected by the sensors 11, 12 into the corresponding proportions.
  • a first sensor 11 can be designed for detecting a current applied to the electric motor 200 and a second sensor 12 for detecting a voltage applied to the electric motor 200.
  • the system 100 further comprises a computing device 20, which is designed to implement an operating point module 21, a trained support vector machine 22 and an output module 23.
  • the operating point module 21, the trained support vector machine 22, and the output module 23 are preferably implemented as software modules which are executed by the computing device 20.
  • the computing device 20 can also be designed to implement the evaluation module described above.
  • the computing device 20 can in particular be a processor with an associated memory, a cloud computing platform, a microcontroller, an application-specific integrated circuit, ASIC, and / or the like.
  • the operating point module 21 is designed to generate an operating point in an n-dimensional operating parameter space from the at least two recorded operating parameters (for example voltage and current), where n 2 2.
  • This operating point which, if the recorded operating parameters are current operating parameters, represents a current operating point, can according to the invention may be used to index a current state of the device 200.
  • the trained support vector machine 22 is designed and trained to subdivide the n-dimensional operating parameter space into at least two classification volumes, each of which indicates different states of the device 200.
  • the support vector machine be training data with corresponding labels y i provided that each of the training points i a respective state x (normal condition, failure condition 1 error condition 2, ...) to assign.
  • a first operating parameter x1 is shown in the horizontal axis and indexed (or: represents), for example, a current intensity applied to the electric motor 200.
  • a second operating parameter x2 is plotted on the vertical axis and indexed (or: represents), for example, a voltage applied to the electric motor 200.
  • First training points 61 which for example indicate a normal state of the device 200, are shown in FIG Fig. 2 represented by circles.
  • Second training points 62 which indicate an error state of the device 200, are shown in FIG Fig. 2 represented by diamonds.
  • an optimal straight line, plane or hyperplane 65 is determined, which ideally divides the training points 61, 62 from one another in such a way that there is a maximum distance d1, d2 from the training points 61, 62.
  • the data points 63 shown for both the first training data 61 and the second training data 62, all of which have the smallest possible distance from the optimal straight line, plane or hyperplane 65, are referred to as support vectors, since these would basically suffice. to train the support vector machine.
  • the support vector machine also gets its name from these support vectors.
  • the optimal straight line, plane or hyperplane 65 divides the operating parameter space, which is spanned by the x1-axis and the x2-axis, into a first classification volume 51, which indicates a normal state, and a second classification volume 52, which indicates an error state ,
  • the support vector machine 22 can be designed in accordance with various known variants and developments of support vector machines, for example as a soft margin support vector machine or as a hard margin support vector machine.
  • the trained support vector machine 22 is now able to assign any current operating point, which was formed by the operating point module, to one of the existing classification volumes 51, 52.
  • the methods described here can also be used with more than two classification volumes, for example three or more classification volumes 51, 52, the first classification volume 51 always indicating the normal state and further classification volumes usually classifying different error states. It is also possible that there are several separate classification volumes, each of which indicates different normal states of the device 200, that is to say states which differ in their operating parameters greatly differ, but both are acceptable in the operation of device 200.
  • the output module 23 is designed to determine a state of the device 200 in accordance with the classification volume 51, 52 in which the generated operating point was assigned by the support vector machine, and to output an output signal 71 which is at least the specific state of the device 200 indexed.
  • the output signal 71 may carry a logic zero as information to indicate a normal state of the device 200 and a logic one to indicate an error state of the device 200.
  • the output signal 71 can carry, for example, binary-coded numerical information which uniquely identifies and characterizes the state of the device 200, for example 0 ("00") - normal state, 1 ("01") - first fault condition, 2 ("10") - second fault condition, and so on.
  • the output module 23 can also be designed to determine a trajectory of the operating point over the course of time in the n-dimensional operating parameter space over a predetermined period of time, or continuously, and to extrapolate its course in the future.
  • the first point in time over time can be determined at which the operating point is likely to assume a position in the n-dimensional operating parameter space which is to be assigned to a different state of the device 200 than the current state.
  • the operating point of the first classification volume 51 which indicates the normal state
  • changes to a position which belongs to the second classification volume 52 which indicates an error state.
  • the indicated error state will occur in X minutes.
  • the output signal 71 can comprise, for example, a time indication which indicates when at least one specific error state will occur, when a change between certain (and which) classification volumes 51, 52 will take place and / or what this change means.
  • the different classification volumes 51, 52 which indicate different error states of the device 200, consecutively indicate more serious error states of the device 200.
  • the extrapolated trajectory can therefore indicate for each class boundary, i.e. optimal straight line, plane or hyperplane 65, which separates two classification volumes 51, 52 from one another, whether and when the operating point will intersect according to the extrapolated trajectory and output a corresponding listing.
  • first error state - operating point is assigned to the second classification volume 52
  • second error state - operating point is assigned to the third classification volume
  • third error state - operating point is assigned to the fourth classification volume
  • FIG. 12 shows a schematic flow diagram for explaining a method for determining a state of a device 200 according to an embodiment of the second aspect of the present invention.
  • the procedure according to Fig. 3 is particular executable by means of the system according to the first aspect of the present invention, in particular with the system 100 according to Fig. 1 ,
  • a step S10 the device 200 is operated.
  • an electric motor or a pump is switched on or the like.
  • a step S20 the at least two operating parameters x1, x2 of the device 200 are detected during their operation, for example as described above with reference to the detection device 10.
  • an operating parameter x1, x2 of the device 200 can be detected by a corresponding sensor 11, 12.
  • an operating point is generated in an n-dimensional operating parameter space based on the at least two recorded operating parameters, where n 2 2.
  • the operating point can be generated S30, as explained above with reference to the operating point module 21.
  • the n-dimensional operating parameter space is subdivided into at least two classification volumes 51, 52, which each indicate different states of the device, using a trained support vector machine 22.
  • the division S40 of the n-dimensional operating parameter space can alternatively also be described as generating classification limits 65.
  • the support vector machine 22 can be trained, for example, as described above, so that a method for training a support vector machine for use in a system for determining a state of a device 200 is also provided here.
  • the subdivision S40 of the n-dimensional operating parameter space takes place in such a way that a first classification volume 51 indicates a normal state of the device 200 and a second classification volume 52 indicates an error state of the device 200. It goes without saying that the n-dimensional operating parameter space can also be divided into further classification volumes, which preferably indicate further error states of the device 200.
  • the subdivision S40 of the n-dimensional operating parameter space is carried out, as is customary in the case of support vector machines, by calculation, depending on the dimension n of the operating parameter space, of an optimal straight line, plane or hyperplane 65.
  • the support vector machine can be designed, for example, as a hard margin support vector machine, in particular if the training data used can be separated linearly, that is to say if an (n-1) -dimensional class boundary can be drawn in the n-dimensional operating parameter space , which defines the classification volumes 51, 52 in such a way that each classification volume 51, 52 includes exactly and only the training data 61, 62 assigned to this classification volume 51, 52.
  • the support vector machine can be designed, for example, as a soft margin support vector machine, for example as an L1 soft margin Support vector machine or an L2 support vector machine.
  • a step S50 it is determined by the support vector machine 22 that the generated operating point is assigned to one of the classification volumes 51, 52. This is usually done using a so-called decision function.
  • a state of the device 200 is determined in accordance with the classification volume 51, 52 in which the generated operating point is arranged.
  • an output signal 71 is generated and output, which at least indicates the specific state of the device 200.
  • the output signal 71 may contain further information, for example an expected time at which the current operating point will change from the classification volume in which it is currently located to another classification volume, and the like.
  • the method can also be used to determine the influences of the various operating parameters on the error states of the device 200 in more detail.
  • a respective normal vector is determined for each classification limit (that is, optimal straight line, plane or hyperplane 65) (advantageously standardized to a length of 1), which is perpendicular to the corresponding class limit.
  • the term “normal vector” here does not refer to the normal state of the device 200, but to the vertical arrangement with respect to the class boundary.
  • the individual entries of the normal vector at the various indices of the normal vector provide information about the direction in which the normal vector points in the operating parameter space, so that their absolute amounts provide information about changes in which operating parameters x1, x2 in particular quickly cause a change between classification volumes.
  • a determination of a respective normal vector on each level or hyperplane which separates the first classification volume from a classification volume which indicates an error state of the device can be carried out.
  • step S90 the value of the largest entry in terms of amount of each particular normal vector can then be determined and output, together with the corresponding index value, so that an error condition of the device 200 can be seen in a workshop or in a laboratory, for example when an error is processed. which operating parameters x1, x2 should be checked as a priority to find the cause of the fault condition.
  • Steps S80 and S90 can be carried out, for example, by an evaluation module implemented by computing device 20.
  • steps S80 and S90 can also be carried out by an external evaluation device which is part of the system 100 and can be connected to the computing device 20 of the system.
  • the corresponding operating parameters x1, x2 can advantageously be checked in this order to ensure that the most likely causes of the error are checked first. In this way, the time required to find the causes of errors can be shortened considerably.
  • This procedure is particularly advantageous if the n-dimensional operating parameter space is divided into at least three classification volumes by the trained support vector machine 22, the first classification volume 51 indicating the normal state of the device and that second classification volume 52 and a third classification volume indicate different error states of the device 200.
  • step S80 a determination of a respective normal vector on each level or hyperplane 65, which separates the first classification volume 51 from a classification volume 52, which indicates an error state of the device 200, can take place.
  • the normal vectors thus indicate on the basis of which operating parameters x1, x2 the operating point from the first classification volume (normal state) would in each case pass into one of the classification volumes 52, which indicates an error state of the device 200.
  • the detected operating parameters are each recorded as portions (in particular as fractions) of a corresponding maximum operating parameter value, the respective maximum operating parameter values being determined in advance and in the detection device 10 and / or Computing device 20 can be stored.
  • the operating parameters are advantageously recorded as values between 0 and 1, both included.
  • Fig. 4 schematically shows a variant of the in Fig. 2 illustrated situation, in which the first training data 61 and the second training data 62 are such that the classification limit (optimal straight line, plane or hyperplane 65, in the present case a straight line) extends parallel to the vertical axis x2.
  • the classification limit optical straight line, plane or hyperplane 65, in the present case a straight line
  • a check of the second operating parameter x2 can be dispensed with entirely.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computational Linguistics (AREA)
  • Medical Informatics (AREA)
  • Automation & Control Theory (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Algebra (AREA)
  • Probability & Statistics with Applications (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
EP18189722.4A 2018-08-20 2018-08-20 Détermination des états d'un dispositif au moyen des machines à vecteur de support Withdrawn EP3614220A1 (fr)

Priority Applications (6)

Application Number Priority Date Filing Date Title
EP18189722.4A EP3614220A1 (fr) 2018-08-20 2018-08-20 Détermination des états d'un dispositif au moyen des machines à vecteur de support
EP19761763.2A EP3811170B1 (fr) 2018-08-20 2019-08-15 Détermination des états d'un dispositif au moyen de machines à vecteur de support
CN201980054387.XA CN112585548B (zh) 2018-08-20 2019-08-15 用于确定设备的状态的系统和方法以及包括该系统的设备
PCT/EP2019/071895 WO2020038815A1 (fr) 2018-08-20 2019-08-15 Détermination des états d'un dispositif au moyen de machines à vecteurs supports
ES19761763T ES2927838T3 (es) 2018-08-20 2019-08-15 Determinación del estado de un dispositivo mediante máquinas de vectores de soporte
US17/270,000 US11244250B2 (en) 2018-08-20 2019-08-15 Determining states of an apparatus using support vector machines

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
EP18189722.4A EP3614220A1 (fr) 2018-08-20 2018-08-20 Détermination des états d'un dispositif au moyen des machines à vecteur de support

Publications (1)

Publication Number Publication Date
EP3614220A1 true EP3614220A1 (fr) 2020-02-26

Family

ID=63517650

Family Applications (2)

Application Number Title Priority Date Filing Date
EP18189722.4A Withdrawn EP3614220A1 (fr) 2018-08-20 2018-08-20 Détermination des états d'un dispositif au moyen des machines à vecteur de support
EP19761763.2A Active EP3811170B1 (fr) 2018-08-20 2019-08-15 Détermination des états d'un dispositif au moyen de machines à vecteur de support

Family Applications After (1)

Application Number Title Priority Date Filing Date
EP19761763.2A Active EP3811170B1 (fr) 2018-08-20 2019-08-15 Détermination des états d'un dispositif au moyen de machines à vecteur de support

Country Status (5)

Country Link
US (1) US11244250B2 (fr)
EP (2) EP3614220A1 (fr)
CN (1) CN112585548B (fr)
ES (1) ES2927838T3 (fr)
WO (1) WO2020038815A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3974924A1 (fr) * 2020-09-29 2022-03-30 Siemens Energy Global GmbH & Co. KG Procédé d'identification hors ligne et/ou en ligne d'un état d'une machine-outil, d'au moins un de ses outils ou d'au moins une pièce usinée par celle-ci

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117853827B (zh) * 2024-03-07 2024-05-14 安徽省大气探测技术保障中心 大气温室气体监测用采样泵工作状态运行监测系统及方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102008058422A1 (de) * 2008-11-21 2010-05-27 Precitec Itm Gmbh Verfahren und Vorrichtung zur Überwachung eines an einem Werkstück durchzuführenden Laserbearbeitungsvorgangs sowie Laserbearbeitungskopf mit einer derartigen Vorrichtung
US20150293523A1 (en) * 2014-04-15 2015-10-15 Mitsubishi Heavy Industries, Ltd. Machine tool diagnostic method and system
EP3296822A2 (fr) * 2016-09-16 2018-03-21 Honeywell Limited Détection de mésappariements d'usine modèle au moyen du groupement de données de paramètres de modèle pour machines à papier ou d'autres systèmes

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103136587B (zh) * 2013-03-07 2015-12-09 武汉大学 基于支持向量机的配电网运行状态分类识别方法
KR102505454B1 (ko) 2016-04-28 2023-03-03 엘에스일렉트릭(주) 무효전력보상장치의 제어장치 및 그의 제어방법
JP6426667B2 (ja) * 2016-08-10 2018-11-21 三菱重工工作機械株式会社 工作機械の工具の異常検知装置及び方法
US11449046B2 (en) 2016-09-16 2022-09-20 Honeywell Limited Model-plant mismatch detection with support vector machine for cross-directional process behavior monitoring
CN107256003A (zh) 2017-05-27 2017-10-17 四川用联信息技术有限公司 一种模糊支持向量机的制造过程多元质量诊断分类器

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102008058422A1 (de) * 2008-11-21 2010-05-27 Precitec Itm Gmbh Verfahren und Vorrichtung zur Überwachung eines an einem Werkstück durchzuführenden Laserbearbeitungsvorgangs sowie Laserbearbeitungskopf mit einer derartigen Vorrichtung
US20150293523A1 (en) * 2014-04-15 2015-10-15 Mitsubishi Heavy Industries, Ltd. Machine tool diagnostic method and system
EP3296822A2 (fr) * 2016-09-16 2018-03-21 Honeywell Limited Détection de mésappariements d'usine modèle au moyen du groupement de données de paramètres de modèle pour machines à papier ou d'autres systèmes

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
SHIGEO ABE: "Support vector machines for pattern classification", 2010, SPRINGER-VERLAG

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3974924A1 (fr) * 2020-09-29 2022-03-30 Siemens Energy Global GmbH & Co. KG Procédé d'identification hors ligne et/ou en ligne d'un état d'une machine-outil, d'au moins un de ses outils ou d'au moins une pièce usinée par celle-ci
WO2022069308A1 (fr) * 2020-09-29 2022-04-07 Siemens Aktiengesellschaft Procédé pour l'identification hors ligne et/ou en ligne d'un état d'une machine-outil, d'au moins l'un de ses outils ou d'au moins une pièce usinée à l'intérieur de celle-ci
US12038739B2 (en) 2020-09-29 2024-07-16 Siemens Aktiengesellschaft Method for the offline and/or online identification of a state of a machine tool, at least one of its tools or at least one workpiece machined therein

Also Published As

Publication number Publication date
CN112585548A (zh) 2021-03-30
CN112585548B (zh) 2022-04-19
WO2020038815A1 (fr) 2020-02-27
US11244250B2 (en) 2022-02-08
US20210312335A1 (en) 2021-10-07
EP3811170A1 (fr) 2021-04-28
EP3811170B1 (fr) 2022-07-06
ES2927838T3 (es) 2022-11-11

Similar Documents

Publication Publication Date Title
DE102012208537B4 (de) Verfahren zum Identifizieren einer Grundursache eines Fehlers in einem gewarteten Fahrzeug und zum Durchführen von Korrekturhandlungen
DE102012223393A1 (de) Verfahren und System für die Grundursachenanalyse und Qualitätsüberwachung von Fehlern auf Systemebene
DE102011076780A1 (de) Verfahren und eine Vorrichtung zur Zustandsüberwachung
EP3841366B1 (fr) Procédé et système pour la détermination directe d'un endommagement théorique d'au moins un organe d'un dispositif
DE112012003403T5 (de) Verfahren, Vorrichtung und Computerprogramm zum Erkennen eines Auftretens von Abnormität
DE102012003020A1 (de) Regelsystem zur Drehzahlregelung eines Antriebsmotors
EP3811170B1 (fr) Détermination des états d'un dispositif au moyen de machines à vecteur de support
DE102019124018A1 (de) Verfahren zum Optimieren von Tests von Regelsystemen für automatisierte Fahrdynamiksysteme
EP3641275A1 (fr) Procédé, dispositif et programme informatique destiné au traitement automatique des identificateurs de données
DE112021005667T5 (de) Verfahren und Einrichtung zum Erkennen von Anomalien in einer mechanischen Einrichtung oder einem mechanischen Bauteil
DE102022122597A1 (de) Verfahren zur überwachung des technischen zustands zur frühzeitigen fehlererfassung in hochspannungsbatterien für elektrofahrzeuge
DE102021124389A1 (de) System und verfahren zum kontrollieren einer fahrzeugqualität
DE102018132658A1 (de) Verfahren zur rechnergestützten Auswertung einer Messung einer elektrischen Größe in einem Hochvolt-Bordnetz eines vorgegebenen elektrisch angetriebenen Kraftfahrzeugs
EP3460727A1 (fr) Procédé d'analyse d'un comportement fonctionnel d'un système technique et unité d'évaluation
WO2021089659A1 (fr) Procédé de détermination d'un écart inadmissible du comportement du système d'un dispositif technique par rapport à une plage de valeurs standard
DE102017124589A1 (de) Verfahren und System zum Auswerten eines Betriebszustandes eines Wasserfahrzeugs
EP3151072B1 (fr) Procede et systeme de reconnaissance de defaut et de surveillance pour un composant de machine commande ou regle electroniquement
DE102019125477A1 (de) Zustandsbestimmungsvorrichtung und zustandsbestimmungsverfahren
DE102020107861B4 (de) System zur Verwaltung von Werkzeugwechselzeiten
DE112021002975T5 (de) Werkzeugdiagnosegerät
DE102021104535A1 (de) Verfahren zum Überwachen der Energieversorgung eines Kraftfahrzeugs
EP3379354A1 (fr) Procédé et dispositif d'usinage automatisé de composants de roue dentée
EP3572893B1 (fr) Analyse d'une pluralité de messages de diagnostic d'un système technique à une pluralité de composants
DE102021134294B4 (de) Hybrid-Ansatz zur Schätzung und Überwachung von Reibungs- und Masseparametern in einem elektrischen Antriebssystem
DE112022005413T5 (de) Überwachungssystem

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

AX Request for extension of the european patent

Extension state: BA ME

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20200827